Steps in Developing a Strategy of Somatic Gene Therapy for Murine OI.

D. Rowe, A. Lichtler, I. Kalajzic, D. Visjnic, J. Kalajzic, P. Liu and ML Stover
Dept. Genetics & Devel. Biol., Univ. of CT Health Ctr, Farmington, CT. USA

We envision a strategy for somatic gene therapy for OI that is based on modifying osteoprogenitor cells in vitro such that the activity of the offending collagen allele is inactivated and replaced with a normal collagen allele. Once these cells have been expanded, they are reintroduced into the affected individual and gradually replace the defective OI matrix with normal matrix. The underlying assumption is that OI bone is under a state of rapid turnover that will select for the transplanted normal bone cells because of its inherently more efficient production bone matrix. We have divided the problems that we feel need a solution for developing this strategy into four topic areas.

1: Inactivate the offending allele – Although ribozymes show great promise to discriminate between a normal and mutant transcript, we have explored an alternative approach based on a modified U1snRNA. We have found that the 10 bp domain that binds to a splice donor site can be modified to interact with a sequence in the terminal exon of a test construct and prevent its expression probably by interferring with its polyadenylation. When introduced into cells by either by transient or stable transfection, it can inactive a test gene, e.g. GFP, by 90-95%. The modified U1 gene has the advantage of high levels of expression independent of cell type and a transcript that is active in the nuclear compartment. Thus it might be used in a complementary fashion with a ribozyme that is active in the cytoplasmic compartment. Currently we are determining the ability of the modified U1snRNA to discriminate between a one or two bp mismatch and whether the U1 concept will work when a site downsteam of the polyadenylation signal is targeted. If this proves to be successful, then the strategy could be used to inactivate a mutant RNA by discriminating a polymorphic site in the 3’ noncoding region of the collagen transcript.

2: Design a retrovector to deliver a correcting cDNA that is regulated in an osteoblast appropriate manner – Because retrovectors can integrate a large DNA fragment into the host genome, we have focused on developing methods for obtaining a high transformation rate in primary osteoprogenitor cell culture. Currently we are using VSV pseudotyping to package the retrovectors. This system produces high titers of virus that can be concentrated by centrifugation, is stable to freezing and should transduce virtually any cell type. We are able to transduce >90% of primary MSF cells and they continue to maintain their ability to differentiate into osteoblasts.

Replication incompetent retrovectors still contain an active promoter/enhancer within the 5’ LTR that can influence the activity of an internal promoter or lead to silencing of the construct in intact animals. Self-inactivating (SIN) retrovectors lack an active 5’ LTR in its infectious form and are not susceptible to silencing. We have placed the bone directed Col1A1 promoter driving various marker genes into a modified Rosa retrovector to test the specificity of the construct when introduced by the SIN vector. In both primary osteoblastic cells and in chimeric mice derived from transduced ES cells, the promoter is activated only in differentiated osteoblasts and remains active in the mature mouse. Thus the promoter design appears to show osteoblast regulation, but it still lacks the strength to drive a collagen cDNA to levels sufficient to replace the inactivated collagen allele. Currently we are examining other regulatory regions within the Col1A1 gene that may provide the strength of expression seen from the Col1A1 minigene.

3: Enriching for an early osteoprogenitor cell population in vitro and optimizing methods for their engraftment of bone – We have characterized the stages of differentiation of the bone cell lineage within MSF cultures. Markers of bone cell differentiation have been developed using Col1A1 derived promoter fragments that become active at different stages of development. Recently we have been successful in using a GFP marker gene that demonstrates activation in real time in culture and is easily detected in standard paraformaldehyde fixed and paraffin embedded bone sections. Using these cell culture models and reagents, it appears that lineage maturation can be reversibly inhibited by either noggin or bFGF. In this manner an expanded population of bone precursor cells can be obtained that are restricted to a defined state of cell differentiation but which will initiate GFP fluoresence if they differentiate into a functioning bone cell. These will be ideal donor cells to test methods of bone engraftment.

To develop a model to evaluate osteoprogenitor cell transplantation, we created the Col2.3*tk mouse. When these vigorous and normal appearing mice are treated with ganciclovir, the bone cells lining the periosteum, trabeculae and endosteum are destroyed and the hematopoietic and osteoclast cell population is greatly reduced. Despite the loss of the bone cells, the mice maintain normal cage behavior during 3-4 weeks of treatment. We believe that the bony surfaces should be most receptive for engraftment. Any progenitor cell that expands to form bone will arise from the transplanted cells since the transplanted cell will be resistant to the toxic effect of ganciclovir. Currently the mice bearing the Col2.3*tk transgene are being bred into a scid/beige background so that the mouse will be a universal model for bone cell transplantation.

4: Sensitive method to evaluate the success of therapy in an OI mouse – The characterization of the oim mouse presented by Dr. Kalajzic elsewhere in the meeting shows the value of a transgene reflecting the high bone formation rate intrinsic to OI. As bone strength improves and the OI cells are replaced by the corrected cell, the transgene activity will fall. Currently we are developing a chimeric marker gene (CAT-ires-GFP) which would measure total osteoblastic activity in the bone and show the distribution of activated cells.

We hope that the reagents and models developed will be useful in optimizing the various step necessary in a successful strategy of gene therapy in mouse and from these studies a design will be formulated that can be evaluated in man.

Reference: Proceedings of the 7th International Conference on Osteogenesis Imperfecta. Montreal, Canada, 1999.